BBA - Bioenergetics 1860 (2019) 148053

Contents lists available at ScienceDirect

BBA - Bioenergetics

journal homepage: www.elsevier.com/locate/bbabio

Pigment-protein complexes are organized into stable microdomains in T cyanobacterial ⁎ A. Strašková1, G. Steinbach1,2, G. Konert, E. Kotabová, J. Komenda, M. Tichý, R. Kaňa

Institute of , Czech Academy of Sciences, Centre Algatech, Novohradská 237, 379 81 Třeboň, Czech Republic

ARTICLE INFO ABSTRACT

Keywords: Thylakoids are the place of the light-photosynthetic reactions. To gain maximal efficiency, these reactions are conditional to proper pigment-pigment and protein-protein interactions. In higher plants thylakoids, the inter- Confocal microscopy actions lead to a lateral asymmetry in localization of protein complexes (i.e. granal/stromal thylakoids) that Photosystems have been defined as a domain-like structures characteristic by different biochemical composition andfunction membrane (Albertsson P-Å. 2001,Trends Plant Science 6: 349–354). We explored this complex organization of thylakoid Membrane heterogeneity Microdomains pigment-proteins at single cell level in the cyanobacterium Synechocystis sp. PCC 6803. Our 3D confocal images captured heterogeneous distribution of all main photosynthetic pigment-protein complexes (PPCs), Photosystem Population heterogeneity I (fluorescently tagged by YFP), Photosystem II and Phycobilisomes. The acquired images depicted cyano- bacterial thylakoid membrane as a stable, mosaic-like structure formed by microdomains (MDs). These micro- compartments are of sub-micrometer in sizes (~0.5–1.5 μm), typical by particular PPCs ratios and importantly without full segregation of observed complexes. The most prevailing MD is represented by MD with high Photosystem I content which allows also partial separation of Photosystems like in higher plants thylakoids. We assume that MDs stability (in minutes) provides optimal conditions for efficient excitation/electron transfer. The cyanobacterial MDs thus define thylakoid membrane organization as a system controlled by co-localization of three main PPCs leading to formation of thylakoid membrane mosaic. This organization might represent evo- lutional and functional precursor for the granal/stromal spatial heterogeneity in photosystems that is typical for higher plant thylakoids.

1. Introduction represents other bioenergetics organelle of higher plants and . Light-photosynthetic reactions proceed on thylakoid mem- Biological membranes were originally described as a fluid mosaic brane that is heterogeneously structured into stacked and unstacked with uniform distribution of proteins and lipids [1]. Later, hetero- regions defined as granal and stromal thylakoids (see e.g. themost geneous membrane areas were found in a form of lipid rafts in animal recent reviews [16–18]). Grana are stabilized by physicochemical cells [2], various bacterial microdomains (MDs) [3–5] or raft-like MDs forces [19,20] controlled by ion compartmentation [21]. All primary in mitochondria [6]; it led to a change in the paradigm of membrane photochemical reactions, including light-harvesting, charge separation organization proposing mosaic macrostructure of cellular membranes and subsequent electron transport processes are catalyzed by hetero- with specific MDs [7]. Recently, a heterogeneous distribution of geneously distributed membrane proteins complexes of Photosystem I membrane proteins has been intensively discussed also in plant cyto- (PSI), Photosystem II (PSII) and cytochrome b6f complex. There is a plasmic membranes [8,9] or in plant mitochondrial membrane [10]. clear lateral asymmetry between granal/stromal thylakoids that form The mitochondrial enzymes of oxidative phosphorylation (OXPHOS) (micro)domains with different biochemical composition and function are known to form heterogeneous membrane compartments [11–13] [22]. Typically, a higher PSII/PSI ratio is typical for granal (stacked) that highly restrict the diffusion of OXPHOS enzymes [14]. Therefore, a and lower for stromal (unstacked) thylakoids [22]. In higher plants, the “plasticity model” of inner mitochondria membranes has been sug- light-harvesting efficiency of photosystems is increased by their mem- gested [15]. brane embedded pigment-protein antennas of photosystems (e.g.

⁎ Corresponding author. E-mail address: [email protected] (R. Kaňa). 1 A. Strašková and G. Steinbach contributed equally to this study. 2 Present address: Institute of Biophysics, Biological Research Center, Szeged, Hungary. https://doi.org/10.1016/j.bbabio.2019.07.008 Received 5 December 2018; Received in revised form 28 June 2019; Accepted 18 July 2019 Available online 22 July 2019 0005-2728/ © 2019 Published by Elsevier B.V. A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053

LHCII). PsaF-YFP fusion at the C-terminus of PsaF, pUC18:psaF-YFP-CmR It contrasts with cyanobacteria in which the light-harvesting an- plasmid was used to transform the Synechocystis WT strain as described tenna complexes, Phycobilisomes (PBS)[23] are situated on the thy- in [40]. This plasmid was constructed by insertion of the YFP coding lakoid membrane surface and can form huge functional megacomplexes sequence (corresponding to a Venus sequence, [43]) at the end of the 3′ with both photosystems [24]. In fact, cyanobacteria represent an evo- coding region of the psaF gene (sll0819). As a selectable marker lutionary ancestor of plant [25,26]. Their thylakoids chloramphenicol-resistance gene (CmR) was inserted downstream of the harbor components of respiratory electron transport chain [27] and psaJ (sml0008) gene which is located downstream the psaF gene in the their thylakoids show no signs of the membrane stacking [28,29]. This Synechocystis genome. Transformants were selected on BG11 plates makes cyanobacteria a unique model system to study localization of with 5 μg/mL chloramphenicol, genome copies were segregated by thylakoid protein complexes independently of membrane stacking. In plating on the chloramphenicol concentration up to 40 μg/mL. PCR was vivo microscopic data have suggested a heterogeneous distribution of used to show integration of YFP and elimination of the WT gene copies. pigment-protein complexes (PPCsincluding PSI, PSII and PBS) in thy- The strains were cultivated on a rotary shaker under moderate light lakoids of various cyanobacterial strains including Synechocystis sp., conditions (white light, 40 μmol of photons m−2 s−1, 30 °C) in liquid Synechococcus sp. or Anabaena sp. Based on these data, several distinct BG11 medium. models of PPCs organization in cyanobacterial thylakoids were built Accession Numbers: Sequence data from this article can be found in proposing either radial (i.e. with variability between inner/outer thy- the GenBank/EMBL databases under the following accession numbers: lakoid layers [30–32]) or lateral [33,34] heterogeneity in PPCs, espe- PsaF (Sll0819), BAA18108; and PsaJ (Sml0008). cially photosystems, composition. Different methods used in these studies also led to different conclusions in respect to the photosystems 2.2. Analysis of protein complexes location: PSI has been preferentially localized either to the outermost (see electron microscopic data in [32]) or to the inner membrane thy- Thylakoid membranes were prepared by breaking cells with zir- lakoids (see hyperspectral confocal fluorescence image data in[31]). konia/silica beads using Mini-Beadbeater (BioSpec Products, USA) as Moreover, in vitro AFM experiments with isolated thylakoid membrane described by [44]. The protein composition of cyanobacterial mem- proposed existence of a specific type of small PSI MDs in Synechocystis branes was analyzed by two-dimensional polyacrylamide gel electro- sp.PCC 6803 [35]; electron microscopy pictures showed arrays of PSII phoresis (PAGE) combining clear native electrophoresis (CN-PAGE) in the same organism [36]; a specific bioenergetics MDs were re- with denaturing SDS-PAGE. CN-PAGE was performed in 4–14% gra- cognized by confocal microscopy in violaceus [37], a pri- dient polyacrylamide gel (acrylamide to BIS-acrylamide ratio was 1:60) mitive thylakoid-less cyanobacterium [38]. These results clearly according to [45] with modifications described in [46]. Native gels showed non-existence of a conclusive model for the thylakoid mem- were photographed and scanned for fluorescence. In- brane organization of PPCs in cyanobacteria. dividual proteins in membrane complexes separated by CN-PAGE were In the present work we addressed complex organization of thylakoid resolved in the second dimension by SDS-PAGE in a denaturing 12–20% membrane proteins by a simultaneous in vivo detection of all major linear gradient polyacrylamide gel containing 7 M urea. Separated PPCs (PSI, PSII, PBS) by means of 3D confocal imaging. Our data proved proteins were further visualized by staining with Coomassie Blue heterogeneous organization of these PPCs into microdomains (MDs) (Quick Coomassie Stain; Generon, Ltd., GB) and the identity of the that define mosaic like structure of thylakoid membrane of stained PSI proteins was verified by mass spectrometry as described in Synechocystis sp. PCC 6803 strain. This conclusion is based on following [47]. methods and results: (1) Simultaneous three channel localization of PPCs [39] has been adapted for 3D imaging of the whole cyanobacterial 2.3. Measurements of PSI activity thylakoid; (2) YFP tagging of PSI [40] allowed localization of PSI in 3D structure of cyanobacterial thylakoid; (3) The new method of image PSI activity was determined from P700 oxidation and re-reduction processing (segmentation) defined/quantified seven possible types of kinetics using a Dual-PAM-100 (Walz, Germany) in the dual-wave- thylakoid membrane MDs; (4) Each MD was typical by particular PPCs length mode Δ(I875-I830). Complete P700 oxidation was induced by a ratios and microdomains did not segregated PPCs from each other; (5) 30 ms saturation pulse (I = 10,000 μmol photons m−2 s−1). Steady Only three most dominant MDs were found in the most cells, namely (a) state (P), maximal (Pm, Pm′) and zero (Po) P700 levels were de- PSI dominant MD; (b) PSI & PSII & PBS MD; (c) PSII & PBS dominant termined at defined actinic light intensities stepwise-increasing (loga- MD; (6) The 3D organization of MDs showed mosaic-like organization rithmic increment) from 0 to 825 μmol photons m−2 s−1 with 30 s of the cyanobacterial thylakoids; (7) MDs are of sub-micrometer in size adaptation periods. Quantum yield of photochemical energy conversion (~0.5–1.5 μm); (8) MDs are very stable in a range of several minutes as at PSI [Y(I) = 1-Y(ND)-Y(NA)] and quantum yields of non-photo- shown by FRAP and time-lapse imaging. We suggest that the observed chemical energy dissipation due to donor side limitation [Y(ND) = 1- MDs stability is conditional for efficient excitation/electron transfer. In P700red.] and due to acceptor side limitation [Y(NA) = (Pm-Pm′)/Pm] conclusion, our structural model of cyanobacterial thylakoid mem- were calculated by Dual-PAM software according to [48]. branes is based on localization of all three PPCs (PSI, PSII, PBS), that interact together and synergistically form the thylakoid membrane 2.4. Measurement of PSII variable fluorescence mosaic. The mosaic distribution pattern of MDs in the cyanobacterial thylakoids reminds of the distinct location pattern of PSII and PSI in The maximum quantum yield of PS II photochemistry (FV/FM) and granal/stromal thylakoids of higher plant chloroplasts [22] and might effective PSII antennas size (σPSII) were detected by fast repetition rate represent its evolutionary precursor. fluorescence (FRRF) method [49] by custom-designed FL3500 fluo- rometer (Photon Systems Instruments, Brno, Czech Republic) at 28 °C 2. Methods with dark-adapted samples (10 min). Single-turnover flashes were in- duced by application of a series of 120 of sub-saturating flashlets (1.5 μs 2.1. Growth of cells and strain generation duration) either blue (λ = 463 nm) reflecting PSII antennas size due to chlorophyll absorption, or amber (λ = 590 nm) light reflecting PSII The YFP tagged strain used in the study was derived from the GT-P antennas size of PBS. The single turnover flash was fitted according to variant of the glucose-tolerant strain of Synechocystis sp. PCC 6803 [41] the model of Kolber and co-workers [49] giving the σPSII. Fv/Fm was

(hereafter referred to as Synechocystis 6803) previously described in calculated as (FM-Fo)/Fm (FM and Fo - the maximum and minimum [42] and hereafter referred to as WT. To generate a strain expressing fluorescence in dark adapted state).

2 A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053

2.5. Oxygen evolution, chlorophyll concentrations and absorption spectra The segmentation algorithm compares the relative intensity of the channels pixel-by-pixel (see Supplemental Fig. 3D). If one from the Oxygen evolution normalized to the chlorophyll a content was three channels (IR,IG or IB) is higher by a factor of 1.15 (≫) than the measured at 28 °C using a Clark-type DW2/2 electrode chamber remaining two, then the corresponding color (red/green/blue) is se-

(Hansatech, United Kingdom) in the presence of 5 mM sodium bi- lected. When the relative intensity of two channels (IR ∧ IG,IR ∧ IB or IB carbonate. The rate of gross oxygen evolution was calculated from the ∧ IG) is similar (~, i.e. the fluorescence ratio is in a range (1/1.15; slope of net O2 evolution measured at saturating irradiance 500 μmol 1.15)), but higher by a factor of 1.15 (≫) than the third channel, photons m−2 s−1 (KL1500, SCHOTT, USA) plus the slope of respiratory yellow, magenta or cyan is selected respectively. The ratio 1/1.15 was

O2 uptake measured in the dark just after light exposure. selected after testing all other approximate values and recognized as the Chlorophyll a content was estimated by extraction from cell pellets most optimal for the used microscopic setup; it means the output of with 100% (v/v) methanol at room temperature using Thermo segmentation (i.e. thresholder binary pictures in 7 colors) co-localized Spectronic Helios Epsilon spectrophotometer according to Porra and co- with original colors of RGB pictures as defined by colorimetric standards workers [50]. CIE 1931 for human vision [52]. This algorithm resulted in 7 combi- – nations, that allowed us to segment RGB pixels into 7 different groups that determined basic MDs with characteristic intensity & color com-

2.6. Confocal microscopy binations: (1) Red MD, dominant in PSII ➔ IR ≫ [IB ∧ IG] (2) Green MD – dominant in PSI ➔ IG ≫ [IB ∧ IR]; (3) Blue MD – dominant in PBS ➔ 2.6.1. Image acquisition IB ≫ [IG ∧ IR]; (4) Yellow MD – dominant in PSI & PSII ➔ [IG ~ IR] ≫ IB; Confocal images were acquired by Leica SP8 microscope (Leica (5) Cyan MD – dominant in PSI & PBS ➔ [IG ~ IB] ≫ IR; (6) Magenta MD Microsystems Inc. Wetzlar, Germany) equipped with SuperK EXTREME – dominant in PSII & PBS ➔ [IR ~ IB] ≫ IG; (7) white/grey MD with a white laser source (NKT Photonics A/S, Birkerød, Denmark) and a HC- “similar” (~) intensity in all channels IR ~ IB ~ IG. The volume of PL-APO CS2-63x/1.4 oil immersion objective. The measuring protocol particular MD in the RGB cube-color space reflects number of possible contained 2 repeated Z-stack scans (with 225 nm step in z) to address a pixel color combinations (IR,IB and IG) in particular MD (see Supple- possible effect of image fading. The optical slices (in Z direction) were mental Fig. 3). The algorithm counting the amount of pixels having optimized by a confocal pinhole adjustment to 1 AU, a 64 × 64 pix different color-combinations was analysed for every cell containing (pixel size 60 nm, 48× zoom) pictures were imaged in x-y. The 3 several Z-stack images separately, and the data were used for calcula- channel RGB pictures were collected in 2 scans: (1) simultaneous de- tions from all cells. tection of YFP signal (Ex: 488 nm; Em: 500–550 nm, detected by Leica Hybrid Detector, gain 150%), chlorophyll autofluorescence (Ex: 2.6.4. Image post-processing, 3D visualization 488 nm; Em: 690–790 nm, detected by Leica Hybrid Detector, gain The three dominant MDs present in segmented images were visua- 130%) and transmission images (at 633 nm, transmission PMT detector, lized in 3D using PovRay software (Persistence of Vision Raytracer Pty. gain 319 V); and (2) phycobilisomes autofluorescence (Ex: 633 nm, Em: Ltd. Williamstown, Victoria, Australia, http://povray.org/). Every pixel 645–680 nm, fluorescence PMT detector, gain: 682 V) measurements. of the three most abundant areas (green – PSI, white/grey – PSI-PSII- The two laser lines were provided by the white laser source, 488 nm at PBS and magenta – PSII-PBS) was represented by an ellipsoid with 3% intensity, 633 nm at 0.03% intensity. diameters: 60 nm, 60 nm, 225 nm according to the slice/pixel size for the 3D model. During the reconstruction process the elementary ellip- 2.6.2. Image processing, construction 3D image soids were transformed into a blob-object and finally rendered by Images were exported by the Leica software to standard 24 bit TIFF Persistence of Vision Raytracer program. format and later post processed by ImageJ macros [51] as merged multipage RGB TIFF images. A possible fading caused by Z-stack mea- surement was corrected for each channel independently by two suc- 2.6.5. Characterization of pixel distribution in RGB color space cessive Z-stacks of the same cell. The averaged fading parameter was The overall distribution of PSI, PSII and PBS in cells was calculated from RGB images. The intensities of red, green and blue channels (IR,IG, calculated for the three RGB channel separately (Red channel – fR, IB) in every pixel were characterized by the RGB values transformed Green channel – fG; Blue channel – fB) from fading detected between k into two coordinate of CIE 1931 color space [53] by ImageJ [51]. The two successive Z-stacks in particular image in the stack as fi = Ii/ I i+ k (k – total number of the slices in z-stacks; i – image position in the single more detailed application of RGB color space for data processing of 3 stack) using an ImageJ macro determining changes in the individual channel images of photosynthetic microdomains has been presented by Konert and co-workers [54]. The total distribution of the fluorescence intensities (Ii and Ii+k) between 2 subsequent Z-stacks measurements. The average fading parameters for RGB channels in PSI-YFP cells was represented in 3D plot. (fR = 1.006; fG = 0.989; fB = 1.002 for single scans) were then applied to all images in the 1st stack as a constant correction factor. The raw Z- 2.6.6. Fluorescence Recovery After Photobleaching (FRAP) measurements stack images were processed by ZEN Black software (version 2.1, Carl Mobility measurements of PSI (based on YFP fluorescence) and PSII Zeiss Microscopy GmbH, Germany) to be displayed in 3D. The trans- (based on chlorophyll autofluorescence) was carried out with a laser- parency mode was used (Threshold 13%, Ramp 50%, Maximum 100%) scanning confocal microscope (Zeiss LCM 880, Carl Zeiss Microscopy with spatial parameter for X-Y = 0.06 μm & Z – 0.225 μm respectively GmbH, Germany) equipped with a Zeiss Plan-Apochromat 63× oil (view angle 45°). objective (NA = 1.4). The YFP and chlorophyll fluorescence was ex- cited by Argon Laser laser (488 nm laser line), fluorescence was de- 2.6.3. Image processing – RGB image segmentation into microdomains tected in two channel mode simultaneously, in the 690–735 nm/ The RGB images were segmented based on the relative fluorescence 530–560 nm spectral ranges for chlorophyll/YFP fluorescence respec- intensity in three detected channels (intensity of PSII fluorescence in tively. The imaging was done with following parameters: zoom: 20×; red channel – IR, intensity of PSI-YFP fluorescence in Green channel – pinhole: 64 μm; pixel dwell time: 4.12 μs; sequential imaging: 1 fps IG, intensity of PBS fluorescence in blue channel –IB). The RGB cube framerate at 128 px × 128 px; dichroic mirror: MBS 488/543/633. was divided into 8 volumes (7 representing the MDs and one for Bleaching was induced during 500 ms when the bleached area was set- background, see Supplemental Fig. 3) when the black background (i.e. up to be 500 nm × 500 nm; it resulted in the bleach depth of 50%. The pixels having intensity lower than 100 on the 8-bit scale for all three exported raw data files were processed by Visual Basic macros channels) was excluded from the segmentation. (Microsoft Excel 2010, Microsoft Corp.).

3 A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053

Table 1 Proportional percent representation of individual photosynthetic microdomains in thylakoid membrane of Synechocystis 6803 PSI-YFP. The values were calculated together for all cells, values represent means ± s.d. from all images acquired during Z-stack scanning (n = 200) when all the images were analyzed together as averaged for all cells. The meta-analysis of the whole cell population, with calculation of MDs presence/absence per single cell is presented in 3 including their statistical analysis. The bold values represents values for three most dominant MDs, see Fig. 4.

Microdomains PSI PSI-PSII-PBS PSII-PBS PSII PSI-PSII PBS PSI-PBS

Color Green White/grey Magenta Red Yellow Blue Cyan

52.5 ± 28.0 21.3 ± 13.0 10.0 ± 15.7 6.0 ± 13.1 6.2 ± 6.8 2.8 ± 7.0 1.2 ± 2.9

2.6.7. Analysis of distribution of microdomains in cell population (Supplemental Table 1) by application of G- [55] implemented in R- Distribution of MD count in every cell was explored using RGB software (version 3.4.4, GNU license). We have tested whether the images after segmentation (see Section 2.6.3). Every cell was analyzed occurrence of cells containing only variations of MDs composed from G, separately (number of cells: 39, full Z-stack containing between 8 and W, M or their combinations (see yellow marked combinations in Sup- 12 images) and the distribution of MD count per single cell in the cell plemental Table 1) was present statistically more often than random population was plotted (value 1 - homogeneous cell, value 7 - maxi- MDs distribution in cell population. The theoretical expected count of mally heterogeneous cell). The calculation employed the process of MDs combinations composed of G, W, and M in population (i.e. sta- segmentation (see Section 2.6.3), when each thylakoid membrane pixel tistically their frequency) was 7 (G, W, M, GW, GM, MW, GWM, see was assigned to one of 7 possible types of MDs (Red, Green, Blue, Supplemental Table 1). The expected frequencies were converted to Magenta, Cyan, Yellow, White marked as R, G, B, M, C, Y and W, re- account the size of the cell population sample we measured (n = 39); it spectively). First, MD size in particular cell was calculated as a per- resulted in frequencies expected 0.62 per every combination of MD. The centage of total thylakoid area by ImageJ macro (see Section 2.6.3) and observed number of cells falling into specific top 3 MD categories presented for all cells together as an average MD size per whole cell composed of G, W, M in the all three groups (I, II and III) were counted population (Table 1), the number of MDs in every cells from population to be 30 (Supplemental Table 1) that was used as an input for G-test in was then plotted separately (Fig. 3). During calculation of MD occur- R software. rence in cell population, only MDs with the sizes equal or higher than 10% were included in the summary. This threshold (10%) for MD size was imposed as it reflected an average size of MDs that covers atleast 3. Results 10% of thylakoid membrane area. 3.1. Strain generation and their photosynthetic parameters 2.6.8. Statistical test on significance of MDs distribution in cell population We have statistically tested if the observed occurrence of 3 largest The organization of major PPCs in thylakoid membranes in vivo was domains in the cell population is of statistical significance in compar- studied in a Synechocystis 6803 strain expressing the Venus variant of ison to their random distribution. First, we have calculated the theo- yellow fluorescent protein (YFP) on C terminus of the PsaF subunit (PSI- retical probability of every possible combination of three largest MDs to YFP strain [40]. This modification of PsaF resulted in the formation of appear in the cell. This included also cells containing only one or two PSI complexes tagged at the cytoplasmic side of the thylakoid mem- MDs that fulfilled requirements of covering more than 10% thylakoid brane (Fig. 1A). The presence of YFP-tagged PSI complexes was con- area (see definition in Section 2.6.7). The expected random distribution firmed by analysis of complexes by clear native (CN) of MDs in cells population was then compared with observed dis- and SDS polyacrylamide gel electrophoresis (2D-CN/SDS PAGE) [46]. tribution based on the analysis of total MDs occurrence in the cell po- Both WT and PSI-YFP showed a similar pattern of PSI and PSII com- pulation (see Section 2.6.7. from Materials and methods). In details, the plexes (Supplemental Fig. 1, 1D scan and 1D fluorescence) in the CN theoretical random distribution (all possible types of cells) included all gel, the only difference represented by low abundance additional green possible combination was calculated based on combinatorics equation: band with mobility similar to the PSII monomer that lacked both PsaF md! and PsaF-YFP (see SDS page in Supplemental Fig. 1). Importantly, SDS N = k!()! md k PAGE (Supplemental Fig. 1) confirmed substitution of PsaF in the PSI- related bands in the in YFP-PSI strain, by a larger PsaF-YFP protein as it where N represented total theoretical combinations of MDs (N), md – all has been previously identified [40]. possible types of MDs (i.e. R, G, B, M, C, W in total md = 7) and k The basic physiological parameters (growth rate or pigment com- stands for count of MDs in a theoretical cell (1, 2 or 3) as written in position) were not affected in the PSI-YFP cells and PSI/PSII ratio isnot Supplemental Table 1. Based on k value, there were three theoretical changed in the Synechocystis sp. PCC 6803 expressing the fluorescently groups: Group I - cells with only one type of MDs (i.e. homogeneous tagged PSI [40]. We further characterized the photosynthetic para- cells), combinations of 1 element (k = 1) from md = 7; Group II - cells meters of this strain (Fig. 1). We have proved no effect of YFP tagging with two types of MDs (two colors), combination of 2 elements (k = 2) on PSII activity; the rate of oxygen evolution and maximal efficiency of from md = 7; Group III - cells with 3 types of MDs (three colors), PSII photochemistry also remained unaffected (Fig. 1B). The YFP tag- combinations of 3 element (k = 3) from md = 7. The sum of these three ging did not affect either PSII connectivity (see legend of Fig. 1) or the equations resulted in total 63 possible theoretical MD combinations (7, effective chlorophyll antenna size of PSII (seeσPSII for blue excitation in 21 and 35 for group I, II and III respectively). This number represented Fig. 1B). Importantly, the YFP tagging, even though YFP sticks out of the theoretical random distribution of MDs with no special preference, the membrane by about 4 nm (see the scheme in the Fig. 1A), its pre- and every possible MDs combination (see Supplemental Table 1) had a sence does not affect interaction of Phycobilisomes (PBS) with PSII as theoretical probability to occur 1/63 = 1.587%. Statistically, the ex- shown in unchanged σPSII for excitation to (see σPSII for pected frequencies between each of 63 possible top 3 MDs compositions amber excitation, Fig. 1D). Similarly, YFP tagging did not affect PSI were 1:1. photochemistry, either in its P700 oxidation–reduction kinetics (Fig. 1C, Further, we have evaluated if the observed frequencies of MDs upper panel), or in the light-dependency of steady-state values of based on experimental data (Supplemental Table 1) differed sig- quantum yields of PSI photochemical energy conversion [Y(I)] and non- nificantly in comparison to the expected theoretical proportions photochemical energy dissipation [Y(NA)] (Fig. 1C, lower panel). As

4 A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053

Fig. 1. Basic physiological characteristics of the PSI-YFP Synechocystis 6803 strain. (A) The structural model of cyanobacterial PSI monomer with YFP (in yellow) attached to the C-terminus of the PsaF subunit (in blue) and exposed to the cytoplasm, the averaged width of thylakoid membrane bilayer (4 nm) and YFP protein are marked; (B) Photochemical properties of PSII in WT (hatched bar) and PSI-YFP mutant (grey bar) cells characterized by measurement of gross oxygen evolution (upper graph) and maximum quantum yield of PSII photochemistry (Fv/Fm). (C) Photochemical properties of PSI characterized by measurement of P700 oxidation + and reduction kinetics (P700 ) and quantum yield of photochemical energy conversion and non-photochemical energy dissipation in WT (black) and PSI-YFP mutant cells (red). Quantum yield of photochemical energy conversion in PSI [Y(I); closed symbols] and quantum yield of non-photochemical energy dissipation due to acceptor side limitation [Y(NA); open symbols] were derived from steady state (P), maximal (Pm, Pm′) and minimal (Po) P700 levels at defined light intensities during rapid light curves with 30 s adaptation periods at stepwise-increased (logarithmic increment) actinic light intensities from 0 to 825 μmol photons·m−2·s−1. No donor side limitation was detected [Y(ND) ≈ 0]. (D) Measurement of chlorophyll (blue excitation) and phycobilisome (amber excitation) effective antenna size of PSII (σ) in WT (hatched bar) and in PSI-YFP cells (grey bar). The blue excitation wavelength was 463 nm and the amber excitation wavelength was 590 nm. The connectivity of PSII reaction centers was p = 0.373 ± 0.007 in WT and 0.410 ± 0.001 in PSI-YFP mutant (blue excitation at 463 nm). All datasets above represent mean ± s.d. of three biological replicates. Organization of photosynthetic protein complexes in WT and in PSI-YFP strains (clear native SDS-PAGE) are presented in the supplemental Fig. 1. the YFP did not affect either PSI/PSII photochemistry or PBS attach- proteins. The RGB images of the PSI-YFP cells acquired in the middle ment to PSII, it makes the PSI-YFP strain ideal model to the study or- cell layer (Fig. 2A) reflected spatial organization and ratios of in- ganization of all PPCs in vivo. dividual PPCs inside thylakoid membranes (Fig. 2). These membrane PPCs were found to be organized into special microcompartments characterized by their RGB color (see RGB color coding description in 3.2. Presence of specific thylakoid membrane microdomains in Fig. 2). The presence of heterogeneous microcompartments in RGB was Synechocystis 6803 detectable independently of confocal microscopy setup (data not shown). In our typical setup (see Materials and Methods), the acquired The native organization of PPCs, namely PSI, PSII and PBS, in thy- images (Fig. 2) showed three dominant colors occurring decreasingly in lakoid membrane was addressed by 3 channel confocal microscopy the following order: green (see e.g. cell C1), white/grey (see e.g. A5) based on detection of PSII (red channel) and PBS (blue channel) auto- and magenta (see e.g. A4). The rest of the basic RGB color combinations fluorescence simultaneously with YFP fluorescence of PSI (green were almost absent (e.g. red) or visible in a few specific cells (see e.g. channel) (Supplemental Fig. 2). The setup allowed us to describe lo- cell C3 for yellow; cell C6 for cyan and blue, cell C3 for yellow in calization of the most important thylakoid membrane proteins thus Fig. 2A). The three dominant combinations of PPCs were repeatedly providing us a good proxy for the overall organization of thylakoid

5 A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053

Fig. 2. RGB image of thylakoid membrane microdomains formed by three pigment-protein complexes (Photosystem I, Photosystem II and Phycobilisome) and microdomain segmentations. (A) The RGB picture represents three independently acquired channels, autofluorescence of Photosystem II (PSII, in red) and Phycobilisome (PBS, blue) together with YFP fluorescence of Photosystem I (PSI in green) (see Supplemental Fig. 2 for separate channels in greyscale). Allpossible combinations of three measured channels are represented by RGB color space with 7 basic combinations as described in the scheme: (1) Red color – dominant PSII emission; (2) Green color – dominant PSI emission; (3) Blue – dominant PBS emission; (4) Yellow – dominant PSI & PSII emissions with minimal PBS auto- fluorescence; (5) Magenta – dominant PSII & PBS emissions with minimal YFP fluorescence from PSI, (6) Cyan – dominant PSI & PBS emissions withminimalPSII autofluorescence; (7) White/Grey – similar PSI, PSII PBS emissions. The combined image shows typical cross section of 39 independently measured cellsrepresenting one layer of the Z-stack scan of the whole cell. Scale bar 2 μm. (B) RGB image segmentation into different microdomains. The image represents binary RGB image after application of segmentation method (see Materials and Methods, Supplemental Fig. 3 and 4) that defines characteristic areas of MD as defined in the enclosed scheme. visible in different cells in a form of specific microcompartments with with dominant PSII & PBS emissions (much smaller YFP emission from hundreds nanometers in size (Fig. 2). Even though these areas were PSI), green MD is characteristic by dominant YFP emission of PSI (and variable in their shape/position inside of thylakoid, they always kept smaller PSII and PBS fluorescence) and white/grey was typical by ba- similar colors. Based on the qualitative analysis of pictures we have lanced of PSII, PBS and YFP-mediated PSI emissions. proposed that these 3 microcompartments represent a specific photo- synthetic MDs, that are defined by their RGB color, it means bythe specific combinations of PSI(YFP)/PSII/PBS fluorescence intensity. Based on definition of RGB color space, magenta MD represents areas

6 A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053

3.3. Microdomains quantification in thylakoid membrane and their cell-to- cell heterogeneity

We quantified abundancy of photosynthetic MDs by a newly de- veloped method of RGB image segmentation (see Materials and Methods, and Supplemental Fig. 3 and Supplemental Fig. 4) and by statistical meta-analysis of whole cell population (see Materials and Methods, and Fig. 3, Supplemental table 1). The calculation of the average MD composition calculated for all cells together confirmed the dominance of the green PSI-MD, magenta PSII-PBS MD and white/grey PSI-PSII-PBS MD (Table 1). These three most dominant MDs covered more than 80% of thylakoid membrane area calculated as average from all cells. However, the bulk-type of analysis (i.e. MD areas were aver- aged per all cells together) covered the hidden population hetero- geneity clearly visible between cells (Fig. 2). Indeed, except green, magenta and white, also the other colors (red, yellow, blue, cyan in Fig. 2) were occasionally present in few specific cells (see e.g. C3 for yellow; cell C6 for cyan and blue as seen in the Fig. 2). Therefore, we also calculated the typical combination of MD areas per single cells (Fig. 3). First, we have estimated the most common number of MDs per cell (Fig. 3) and later we also calculated the typical MD composition per cell (see Supplemental Table 1). In both cases, the most frequent number of the dominant MDs was three, and 80% of cells were heterogeneous (Fig. 3) with most dominant combination of MDs Fig. 3. Distribution of number of microdomain counts per single cell in the studied population of Synechocystis sp. PCC 6803 cells. The figure characterizes represented by green PSI-MD, magenta PSII-PBS MD and white/grey distribution of microdomain counts per single cell in the population of cyano- PSI-PSII-PBS MD. This was confirmed by both the averaged RGB image bacteria. The value “1” represents homogeneous cell with single type of MD, segmentation (Table 1) and the meta-analysis cell-by-cell for whole values in the range “2” to “7” reflect heterogoneous cells with various number population (see Supplemental Table 1 and Fig. 3). Further, we also of MDs (see Supplemental Table 1 for details). The MDs counts in single cell statistically tested a significance of the three dominant MDs in single have been calculated from 3D cell image after image segmentation (see Section cell (green, magenta, white) by G-test (see Materials and methods, 2.6.3) when only membrane areas with the size equal or higher than 10% were Supplemental Table 1). Indeed, the analysis proved that the observed counted as MDs (see Section 2.6.7). Insert shows number of homogeneous/ most frequent MD combination in the population of cells significantly heterogeneous cells. The bottom table shows total number of particular MDs in differed from MD combinations that assume the equal probability of cell populations (counted cell by cell) where R, G, B, M, C, Y and W mean Red, their occurrence (see Supplemental Table 1). Green, Blue, Magenta, Cyan, Yellow, and White/Grey MDs respectively. Total We subsequently analyzed the three dominant MDs in more details number of cells used for analysis was 39. The microdomains combinations presented in the analysis are based on 3 channel confocal fluorescence mea- (Fig. 4). In contrast to recently described seemingly homogenous small surement (PSI (YFP) fluorescence – green; PSII fluorescence – red; PBS fluor- MDs of PSI observed in isolated thylakoids of Synechocystis 6803 in vitro escence – blue). The resulted RGB color pictures can contain 7 possible types of [35] we found that in our MDs the major photosynthetic complexes are MDs (see insert of Color coding): (1) Red color – MD with dominant PSII not strictly segregated. So, the magenta PSII-PBS MDs always exhibited emission; (2) Green color – MD with dominant PSI emission; (3) Blue – MD with a weak signal of YFP reflecting the presence of small amount of PSIin dominant PBS emission; (4) Yellow – MD with dominant PSI & PSII emissions this MD (Fig. 4A), nevertheless the signal of PSII (red) and PBS (blue) and minimal PBS autofluorescence; (5) Magenta – MD with dominant PSII& fluorescence was almost doubled in comparison with the YFP fluores- PBS emissions with minimal YFP fluorescence from PSI; (6) Cyan – MD with cence of PSI (see e.g. peripheral part of the cell profile in the Fig. 4B). dominant PSI & PBS emissions with minimal PSII autofluorescence; (7) White/ The white/grey MDs were typical by an approximately equal presence Grey – MD with similar PSI, PSII, and PBS emissions. For further details, see of all three PPCs, which was reflected by their balanced fluorescence Supplementary Table 1. emissions (see histogram/profiles in Fig. 4D and E) while in the green MDs the YFP emission of PSI exceeded almost twice the emission of PSII mosaic from different angles (Supplemental Fig. 6A, Supplemental and PBS documenting the dominance of PSI over PSII and PBS in this Fig. 7A, Movie 6 and Movie 7). The 3D model reflects a separation of MD (see fluorescence histogram/profile in Fig. 4G and H). Thus, the PPCs into the magenta PSII-PBS MDs surrounded by white/grey PSI- photosynthetic MDs with specific PSI/PSII/PBS ratios define a mosaic PSII-PBS MDs and more distantly by green PSI MDs. As the white/grey like structure of thylakoid membrane. intermediate PSI-PSII-PBS MD exhibits the balanced fluorescence of all three PPCs (Fig. 4D), it was not so visible in the single layer RGB image 3.4. 3D model of thylakoid microdomains distribution and microdomains (Fig. 3) also due to a limited spatial resolution of confocal microscopy. stability We also tested a stability of MDs using a time-lapse imaging to- gether with Fluorescence After Photobleaching Recovery (FRAP) The mosaic like distribution of MDs within the thylakoids has been method (Fig. 6). The two channel detection allowed us to address PSI further characterized by different Z sectioning and 3D reconstruction of and PSII mobility simultaneously (Fig. 6). A comparison of “Pre-bleach” thylakoid membrane system in Synechocystis 6803 (see Fig. 5 for (t = 0 s), “After-bleach”(t = 6 s) and “Recovery”(t = 300 s after workflow). Size of photosynthetic MDs (particularly green and ma- bleaching period) for (reflecting localization of PSII) and genta) were not completely uniform (between 0.5 and 1.5 μm) and they YFP (reflecting localization of PSI) showed no change in organization of were often separated by sharp edges (see Fig. 5C, see e.g. the enclosed MDs and overall mosaic like structure of thylakoids during 6 min of Movie 5). The white/grey MDs were often situated in between the other image acquisition (Fig. 6A) also including the bleaching period (see two most dominant MDs in the most typical cyanobacterial cells (see cyan arrow in Fig. 6A). Moreover, the MDs outside of bleach area with MDs variability in Fig. 3). This is shown in our model of thylakoid high abundancy of PSI (see green arrows in Fig. 6A) and PSII (see red membrane (Fig. 5D) based on the Z-stack images (Supplemental arrows in Fig. 6A) were also stable during image acquisition on a scale Fig. 6A, Supplemental Fig. 7A), which shows organization of PPCs into

7 A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053

Fig. 4. Proposed pigment-protein composition of photosynthetic microdomains (MDs) based on fluorescence intensities. The model shows proposed pigment-protein complexes composition in three dominant microdomains (magenta – PSII-PBS MDs; white/grey – PSI-PSII-PBS MDs; green – PS MDs) that reflects Photosystem I (PSI), Photosystem II (PSII) and Phycobilisomes (PBS) content and distribution in MDs. The model is based on fluorescence characteristics of MDs represented by histo- grams and fluorescence profiles of selected cells. The three fluorescence intensities detected by confocal microscope are drawn in their characteristic colorsforred channel (chlorophyll emission from PSII), green channel (YFP emission from PSI) and blue channel ( emission from PBS). Model and fluorescence characteristics of magenta MD, i.e. PSII-PBS MD: (A) Typical fluorescence histograms in PSII-PBS MD; (B) Profiles of fluorescence intensity in selected cell in PSII-PBS MD; (C) Simplified model of PSII-PBS MD. Model and fluorescence characteristics of grey/white MD, i.e. PSI-PSII-PBS MD: (D) Typical fluorescence histograms of PSI-PSII-PBS MD; (E) Profiles of fluorescence intensity in selected cell in PSI-PSII-PBS MD; (F) Simplified model of PSI-PSII-PBS MD. Model and fluorescence characteristics of green MD, i.e. PSI MD: (G) Typical fluorescence histograms of PSI MD; (H) Profiles of fluorescence intensity in selected cell in PSI MD; (I) Simplified model of PSI-PSII-PBS MD. Data represent typical histograms from magenta, green and white/grey areas visiblein Fig. 2. of minutes. The FRAP measurement also allowed us a detailed assess- PSI, however there were no quantitative analysis details about PSI co- ment of the PSII and PSI mobility in the membranes (see Fig. 6B). The localization with PSII and PBS and no 3D model of PSI occurrence. data were in line with the previously shown (see e.g. review [56]) low Further, in the work of Casella and co-workers [33] authors have dis- mobility of PSI and PSII in cyanobacteria (see Fig. 6B). Such a low cussed and shown 2D co-localization of PSI + PSII, however PBS were mobility clearly explains the stable mosaics like organization of PPCs again not detected and none from the picture were quantified by image within the thylakoid membrane of Synechocystis 6803 (Fig. 5). segmentation. Finally, in the Vermaas's work [31] authors did indeed address all the three main channels PPCs, however the occurrence was 4. Discussion judged from spectra deconvolution by hyperspectral confocal fluores- cence method without any fluorescence tagging. We used a different, There have been several attempt to study thylakoid membrane more native approach. It is based on a combination of fluorescence heterogeneity in cyanobacteria by various in vitro methods including tagging of protein complexes and life imaging of cyanobacterial cells by proteomics [57], electron microscopy [28,32,36], AFM microscopy confocal microscope. Interestingly, we came with different conclusion [35] or cryo-imaging [34]. Based on the confocal microscopy approach, in contrast to work [31] as we did not see radial but rather lateral MacGregor-Chatwin and coworkers [35] has shown heterogeneity of heterogeneity in PPCs. Further, we have also seen no segregated

8 A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053

Fig. 5. Thylakoid membrane mosaic of photosynthetic microdomains (MDs) based on 3D images in Synechocystis 6803 PSI-YFP cells. MDs are formed by pigment- protein complexes of Photosystem I (green), Photosystem II (red) and Phycobilisome (blue) but only 3 of the 7 possible combinations are dominant (see yellow rectangle): (1) Green MD (with prevailing PSI abundancy); (2) white/grey MD (balanced PSI–PSII–PBS content); (3) Magenta MD (prevailing PSII-PBS abundancy); (4) cyan (prevailing PSI–PBS abundancy); (5) yellow (prevailing PSI–PSII abundancy); (6) blue (prevailing PBS abundancy); (7) red (prevailing PSII abundancy). (A) Z-stack images of a single cell (see Movie 1); (B), A reconstituted 3D Z-stack of single cell from Panel A, a distance between section was 0.225 μm; (C), MDs structure and organization in different cells constructed based on Z-stack images (Supplemental Fig. 5, Movies 2, 3, 4 for cells 1–3);(D) A 3D model of MDs, constructed from cell no. 6 (see Supplemental Fig. 6 for workflow and Movie 6), based on segmented data (see example in Supplemental Fig.4). microdomains of PSI in vivo in contrast to conclusion of MacGregor and (dominant PSII & PBS emissions), green MD (dominant YFP emission of co-workers [35]. Based on our data, PSI complex in thylakoid mem- PSI) and white/grey MD (balanced PSII, PBS and YFP-PSI emissions). brane is every time inter-mix with PSII, with different PSI/PSII ratio in This represents the most common type of MDs organization that we different MDs (e.g. PSI content is very high in green MD,see Fig. 4). further discuss in our model (Fig. 7). MDs organization was different We suggests that thylakoid membrane organization can be consider only in some specific cells (see e.g. cell C3 for yellow; cell C6 forcyan as a specific system of 3 PPCs that works together and form thylakoid and blue, cell C3 for yellow in Fig. 2) and this can be easily explained by membrane mosaic (see Fig. 5). In fact, there is only a limited combi- the observed population heterogeneity (see Supplemental Table 1 and nation in PPCs as we have described recently by new photosynthetic Fig. 3). Such a heterogeneity is well known phenomenon and con- parameter called Protein arrangement factor (see Supplemental Fig. 8 sidered as fundamental property of cellular systems [58,59] including and [54]). The authors proved that even during long-term irradiation, bacteria [60]. Similarly, this heterogeneity within bacterial populations the fluorescence ratios of PSI/PSII/PBS (defining MDs structure) didnot is well defined and provides a mechanism to increase range of responses change on average on cell population level [54]. Our new approach to changing environmental conditions [61]. In fact, deviations of a presented here allowed us to further identify the internal heterogeneous particular parameter (e.g. in our case “MDs count per single cell”, see organization of thylakoid membrane as a complex system defined by Fig. 3) from the average in the whole population (in our case, “the three PPCs. We describe cyanobacterial thylakoid membrane as a mo- averaged MDs size per all cell”, see Table 1) are the basis of Darwin's saic structure consisting of stable photosynthetic MDs (Fig. 5) typical by theory of evolution by natural selection [62]. Therefore, our data point specific ratio of three major PPCs: PSI, PSII andPBS(Fig. 4). In addi- out to the necessity to consider population variability in cyanobacteria tion, we performed the meta-analysis of data cell-by-cell and this also in future studies addressing single cells (e.g. microscopy pictures of confirmed the presence of several MDs in a single cyanobacterial cell single cell). Our data indicates that different cyanobacterial cells of the (Fig. 3). We have analyzed number of MDs per single cell, the analysis same species are not “uniform units” and cells in the population can proved that only 20% cells were relatively homogenous (see insert in differ significantly (see e.g. observed variability inthe Fig. 2, totally Fig. 3) and the remaining cells were heterogeneous with MDs. These green cell C1 versus “magenta cell” A4). To overcome this natural po- cells had usually three major types of MDs represented by magenta MD pulation variability, a higher number of cells (tens or hundreds) needs

9 A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053

Fig. 6. Time lapse imaging of PSI and PSII fluorescence of representative FRAP image sequence of Synechocystis 6803 PSI-YFP cells. (A) FRAP image sequence cells acquired for PSI (green) and PSII (red), for pre-bleach (“0 s”), after-bleach (“6 s”) and after fluorescence recovery (“300 s”). The green/red arrows show MDswith high PSI & PSII signal. Cyan arrows and circle indicates the position of the bleached area. (B). The averaged kinetics of PSI and PSII fluorescence in the bleached area. Data were corrected for fluorescence changes in the bleached cell and normalized to pre-bleach values (data represent mean and±s.d.for n = 8). Scale bar 1.5 μm. to be compared if we want to make robust conclusions for whole cell 4.2. Stability of photosynthetic MDs population. The limited types of hues observed in the most cells (compare green/magenta/grey hues in the Figs. 2, 5) reflects an existence of only 4.1. Separation of photosystems into the distinct microdomains particular combinations of PSI, PSII and PBS within thylakoids; this can be also quantified numerically by calculation of chromaticity values of Mosaic like structure of the cyanobacterial thylakoids reflects a thylakoid membrane (see Supplemental Fig. 8). The method can loca- partial separation (better say different distribution) of both photo- lize all thylakoid pixels into CIE1931 color space (for details see [39]). systems. In higher plants, PSI and PSII are differently distributed be- The analysis showed that 90% of thylakoid membrane pixels covered tween stromal/granal parts of thylakoid [22]. This heterogeneous lo- only 8.9% area of the CIE1931 color space (see Supplemental Fig. 8 and calizations of photosystems in cyanobacteria are clearly visible in our [54] for details on the method). It means that only certain combinations images showing a high contrast between green and magenta MDs of fluorescence signal of PSI, PSII and PBS (i.e. their concentration ra- highly enriched in PSI and PSII, respectively (see Fig. 4 for details). The tios) are present in the thylakoid membrane mosaic. The presence of “intermediate” white/grey domain with balanced PSI, PSII and PBS (see photosystems in the particular photosynthetic MD (Fig. 5) also limits Fig. 4 for details) could represent an area with recently described su- their free diffusion in the membrane. The stability of photosynthetic percomplexes of all these PPCs [24]. MDs has been confirmed by FRAP and time lapse imaging (Fig. 6). This Our data newly describes model of PSI/PSII co-localization in cya- stability is closely related to the observed low mobility of PSII and PSI nobacteria. Up to now, there has been no conclusive model of PSI/PSII (Fig. 6) which is in line with previous reports (see e.g. reviews in separation in the cyanobacterial thylakoids. Both, a radial hetero- [56,66]). Generally, mobility of photosynthetic proteins is approxi- geneity with difference in PSI/PSII between inner and outer thylakoid mately 10–100 times lower than mobility of mitochondrial OXPHOS layers [31,32]), and a lateral heterogeneity due to variability in PSI/ enzymes that have a similar size as the cyanobacterial PPCs based on PSII ratio within the single thylakoid membrane layer [33,34], have FRAP halftime (compare approximate 100 s with 1–5 s half time for been proposed. Due to the obvious spatial restrictions of light confocal PPCs and OXOPHOS proteins respectively, see [14,67] or reviews microscopy (including hyperspectral confocal fluorescence imaging [56,68]). Both membranes, inner mitochondrial [69,70] and thylakoid [63]) the standard methods cannot resolve radial distribution of pho- membrane [71] are highly crowded with proteins and have similar li- tosystems between membrane layers with the sufficient resolution. The pid:proteins ratio around 20:80. It indicates that restricted mobility of PPCs composition of the MDs we observed thus represents an average PPCs is not solely due to a proteins crowding as proposed previously PPCs composition of multiple thylakoid membrane layers. We can only [72] but an unknown factor(s) may limit their mobility. In analogy with speculate that due to the steric hindrance of PSI, preventing PSI other biological membranes (see e.g. review of [7]), the unknown factor movement into inner membrane layers, the PSI supercomplex can only could be protein-protein interactions within the membrane [73], cy- be present in the most outer/inner surface thylakoid layers as described toskeleton corralling [74] or by lipids organization into nanodomains for granal thylakoids of higher plants [22]. On the other hand, the [2]. The photosynthetic MDs thus represent rather a stable, a skeleton- lateral heterogeneity of photosystems in cyanobacteria is more than like structure with moderate dynamics during long-term (minutes) ac- obvious from our confocal microscopy data (see Fig. 5) and it is also in climations at specific conditions (e.g. “mobilization” of proteins byvery line with the previous data based on cryoimaging detection of PSI/PSII intense "red" line, see [33,75,76]) or they can change faster at sub- heterogeneity [34]. We do not know driving force for this spatial PSI/ micrometer scale [77]. PSII separation but it can be some structural consequence of protein- The restricted mobility of photosystems has to affect redistribution protein and lipid-protein interactions [64], an interplay between phy- of light between photosystems during the state transition process which sicochemical forces [19,20] controlled by ion compartmentation [21] mechanism remains still unclear in cyanobacteria (see e.g. review or some specific protein clustering [65]. [78]). Indeed, the observed stability of MDs (Fig. 6) together with low abundance of the specific PSI-PBS MDs (Table 1) excludes any

10 A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053

(caption on next page)

11 A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053

Fig. 7. Model of thylakoid membrane mosaic formed by photosynthetic microdomains (MDs) in Synechocystis 6803. The cyanobacterial model is based on our confocal image data, model of the plant thylakoid membranes was taken from Albertsson [22]. MDs are formed by photosynthetic pigment-proteins, Photosystem I (PSI – green channel), Photosystem II (PSII – red channel) and Phycobilisome (PBS – blue channel), see RGB color coding scheme. (A) Bright field image of a single Synechocystis 6803 PSI-YFP cell with visible cytoplasmic membrane, scale bar 0.5 μm. (B), Fluorescence image of the same Synechocystis 6803 PSI-YFP cell from the panel A. Picture shows localization of dominant photosynthetic MDs: (1) PSI MD in green; (2) PSI-PSII-PBS MD in white/grey; (3) PSII-PBS in magenta. Scale bar (0.5 μm) reflects an approximate size of single magenta MD. (C) Proposed model of MDs arrangement in 3 bilayer of Synechocystis 6803 TMs (Top and Side view) based on experiment (Panel A) the 3D model of MDs (see Supplemental Fig. 6, Movie S6). The used colors represent dominant MDs (green, white/grey and magenta) with defined combinations of PPCs (Fig. 4); we have to note that PBS is situated on TM surface in all types of MDs. Scale bar 0.5 μm. (D) Typical MDs-like organization in higher plants thylakoid that are arranged into granal and stromal membranes with margins in between. Granal thylakoid membranes contain mostly PSII with plant light-harvesting antenna proteins (resembling magenta PSII-PBS MD from cyanobacteria), stromal thylakoid membranes contains mostly PSI (resembling magenta PSII-PBS MD from cyanobacteria), and grana margins represent a mix of both (resembling white/grey PSI-PSII-PBS MD from cyanobacteria). mechanisms requiring a long-distance movement of PPCs (including 4.4. Properties of photosynthetic microdomains in comparison to other PBS) as it has been proposed previously [79,80] and also discussed for membrane microcompartments of (cyano)bacteria and organelles [81], other photothrophs contaning PBS on membrane sur- face. On the other hand, this implies a mechanism of state transition The newly described photosynthetic MDs in Synechocystis 6803 relying on just a small PPCs reorganization [82] that can be further thylakoids have different characteristics in comparison to bacterial affected by membrane fluidity [83], PSI trimerization [84], phycobili- [107] and mitochondrial [11–13] bioenergetic MDs that are less stable somes decoupling [85–88] or by direct excitation quenching in PSII and smaller. The photosynthetic MDs are much bigger (between 0.5 and [88]. 1.5 μm) in comparison to any other membrane microcompartments in cyanobacteria described before (compare 0.1–0.2 μm size of small patches of FtsH proteases [108]); bioenergetics MDs in Gloeobacter 4.3. Function of the separated stable PSII and PSI microdomains violaceus [37]; MDs of OXPHOS complexes [109]; nanodomains con- taining only PSI [35]; zones with CurT protein [110]; PratA-defined We propose that location of PPCs (Figs. 2 and 4) in distinct MDs regions of proteins biogenesis (see e.g. [111] and others see review plays an important role in the regulation of electron and light energy [112]). The above mentioned “nanodomains” were defined based on transfer processes [89]. The low mobility of PPCs seems to be a pre- tagging of a single protein and determination of its location in a requisite for the efficient exciton energy transfer [90] as also proposed membrane region with a specialized function (e.g. biogenesis of pho- previously for light-harvesting process in higher plant thylakoids tosystems, respiration etc.). On the contrary, the photosynthetic MDs [91,92]. In higher plants, the efficient transfer of excitations is allowed reflect co-localization of three main pigment-protein complexes inthe by a close interaction of either PSI or PSII with their membrane antenna thylakoids - PSI, PSII and PBS – into a specific mosaic. Our three proteins [93,94]) or by formation of megacomplexes of the antenna channel imaging of these PPCs was thus able to describe a system or- proteins with both photosystems [95,96]. On the higher mesoscopic ganizing these three important protein complexes. Therefore, our level, reorganization of these supercomplexes into disordered/ordered method allowed us to define an overall structural model of thylakoid arrays further affects light harvesting and lateral diffusion processes in membrane for the most typical cyanobacterial cells (Figs. 4, 5 for data plants (see recent reviews [97,98]). In cyanobacterial thylakoids simi- and Fig. 7 for model). Our model describes thylakoid membrane as a larly organized arrays of PPCs have also been found [36] together with mosaic of photosynthetic MDs; these MDs are then formed by specific active megacomplexes consisting of both photosystems with attached ratio of PSI, PSII and PBS, their specific co-localization into MDs then PBS [24]. Indeed, these nanoscale arrays of PPCs are then visible using affects and constrains the overall thylakoid membrane structure. Itis electron transmission or atomic force microscopy (see e.g. [33,36] or plausible, that the other small “nanodomains” (see description above) reviews [99,100]). These arrays of PPCs then form the stable photo- could overlap/interact with our larger photosynthetic MDs; for instance synthetic MDs visible in our 3D model of the most typical organization we might attribute the center of biogenesis to a very low abundant PSII- of cyanobacterial thylakoids, which we observed (Fig. 5). PSI MD (Table 1) as Bečková and co-workers [113] have recently at- The separation of PSII and PSI into the distinct domains can reduce tributed supercomplexes of PSI trimers and PSII monomers to the pro- excitation energy spillover from photochemically less efficient PSII to- cess of photosystem biogenesis. wards PSI as proposed for higher plant thylakoids [101]. We need to note that the model is not valid for all observed cells due to the in- 4.5. Comparison of cyanobacterial microdomains with structure of higher evitable population heterogeneity. About 20% of the cells showed ra- plants thylakoids ther homogenous distribution of PPCs (see Fig. 3), in which the me- chanisms responsible for MD organization of PPCs might not be We suggest that organization of PPCs into photosynthetic MDs in working. The previously described functional nanodomains for plasto- Synechocystis thylakoids represents an evolutionary precursor of orga- quinone diffusion between PSII and the cytochrome b f [102,103] are 6/ nization of higher plant thylakoids into granal/stromal parts (see smaller in comparison to MDs, however it is plausible that they are Fig. 7). This is based on structural (see Fig. 7) and functional analogy as differently distributed between MDs. A precise PSI/PSII organization is discussed in previous paragraphs. The model represents the dominating also needed for regulation of the ratio of photosynthetic cyclic/linear arrangement of thylakoids even within the heterogeneous population of electron transport (see our model comparing plant/cyanobacterial cyanobacterial cells (see MDs variability in Fig. 3 and supplemental thylakoids, Fig. 7) in the most typical cyanobacterial cell (see MDs Table 1). Image segmentation (supplemental Fig. 3) and modelling of variability in Fig. 3) where the PSII and PSI MDs could then represent experimental data (supplemental Figs. 6 and 7) allowed us to re-con- areas with dominancy of linear and cyclic electron flow, respectively, in struct a 3D model of thylakoids (Fig. 5D, supplemental Fig. 6). It shows analogy with higher plant thylakoids [104]. The structure hetero- the localization of three dominant MDs in the thylakoid membrane in geneity in PSII/PSI between MDs can then result in functional hetero- the typical cyanobacterial cells (Fig. 7): a central PSII-PBS MD (in geneity in photochemical processes that is often detected by fluores- magenta) was surrounded concentrically by the PSI-PSII-PBS MD (in cence parameters showing domains with different rates of photo- white/grey) and more distantly by the PSI MD shown in green (Fig. 7). reduction [105,106]. This organization resembles a granal/stromal thylakoid heterogeneity

12 A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053 of higher plant chloroplasts as proposed long time ago [114]. Indeed, Science Foundation and by institutional project Algatech plus (MSMT the thylakoid membrane structure in PBS-less mutants [28] resembles LO1416) and Algamic (CZ.1.05/2.1.00/19.0392) provided by the Czech granal stacks similar to those in the higher plant thylakoids. Interest- Ministry of Education, Youth and Sport. The work od G.S. has been ingly, the major magenta and green MDs have often diameter in the funded by GINOP-2.3.2-15-2016-00001 grant. The authors want to range of 500–1500 nm that is comparable in size with thylakoid grana thank Jaroslav Krafl for his enthusiastic approach in taking confocal of higher plants (see Figs. 3, 5 and our model in Fig. 7) ranging from microscopy pictures of Synechocystis sp. 6803 cells and Eva Prachová 300 to 600 nm (see e.g. [115,116]. Moreover, in analogy to the cya- for her long-term carring about cyanobacterial strains/mutants in the nobacterial membranes (Fig. 4), PSII and PSI in the plant thylakoids are laboratory of photosynthesis. not strictly segregated but only unevenly distributed between granal (with more PSII) and stromal (with more PSI) thylakoids. Therefore, it Abbreviations is tempting to propose, that the evolutionary loss of PBS and appear- ance of membrane-embedded antennas in plants resulted in develop- FRAP Fluorescence after photobleaching recovery ment of a different mechanism of PSI/PSII compartmentation (Fig. 7). MD Microdomain Thus, during evolution the photosynthetic MDs in the unstacked thy- OXPHOS Enzymes of oxidative phosphorylation in mitochondria lakoids of cyanobacteria were substituted by the photosystem com- PBS Phycobilisomes partmentation based on membrane stacking (Fig. 7). This might also be PPCs Pigment-Protein complexes, namely Photosystem I, Photo- connected with evolution of channels/transporters [117]) as the system II and Phycobilisome membrane stacking is driven by physicochemical forces [19,20] and PSI Photosystem I controlled by ion compartmentation [21]. In conclusion, the described PSII Photosystem II photosynthetic MDs in the cyanobacterial thylakoids can be viewed as a RGB Red & Green & Blue colors system representing the 3 channel new example of compartmentation of cellular processes in prokaryotes confocal imaging

(reviewed in [118]) apart from compartmentation of genetic informa- σPSII effective antenna size of photosystem II tion [119], cytoskeletal proteins [5] and bacterial sensory system YFP Yellow Fluorescence Protein [120]. References 5. Conclusions [1] S.J. Singer, G.L. Nicolson, The of the structure of cell mem- We have described heterogeneous organization of all main pigment- branes, Science 175 (1972) 720–731. [2] K. Simons, E. Ikonen, Functional rafts in cell membranes, Nature 387 (1997) protein complexes - PPCs (Photosystem I tagged by YFP, Photosystem II 569–572. and Phycobilisomes) in the thylakoids of Synechocystis 6803. Our 3D [3] K. Matsumoto, H. Hara, I. Fishov, E. Mileykovskaya, V. Norris, The membrane: confocal images depicted heterogeneous distribution of PPCs in thyla- transertion as an organizing principle in membrane heterogeneity, Front. Microbiol. 6 (2015) 21. koid membranes in a form of stable MDs. These microcompartments are [4] L. Shapiro, H.H. McAdams, R. Losick, Generating and exploiting polarity in bac- of sub-micrometer in sizes (~0.5–1.5 μm) and they are stable in the teria, Science 298 (2002) 1942–1946. range of minutes. MDs are defined by particular ratios of PPCs asthey [5] H. Strahl, F. Burmann, L.W. Hamoen, The actin homologue MreB organizes the contain different mix of all studied complexes – photosystems and bacterial , Nat. Commun. 5 (2014). [6] M. Sorice, V. Manganelli, P. Matarrese, A. Tinari, R. Misasi, W. Malorni, phycobilisomes. The most prevailing MD is represented by MD with T. Garofalo, Cardiolipin-enriched raft-like microdomains are essential activating high Photosystem I content (and low PSII and PBS concentrations) that platforms for apoptotic signals on mitochondria, FEBS Lett. 583 (2009) allows partial separation of Photosystems similar to stromal/granal 2447–2450. [7] G.L. Nicolson, The fluid-mosaic model of membrane structure: still relevant to heterogeneity of photosystems in higher plant thylakoids. Therefore, we understanding the structure, function and dynamics of biological membranes after propose that the cyanobacterial MDs might represent evolutional and more than 40 years, Biochim. Biophys. Acta, Biomembr. 1838 (2014) 1451–1466. functional precursor for the granal/stromal heterogeneity in higher [8] J. Malinsky, M. Opekarova, G. Grossmann, W. Tanner, Membrane microdomains, rafts, and detergent-resistant membranes in plants and fungi, in: S.S. Merchant plants thylakoids. (Ed.), Annual Review of Plant Biology, vol 64, Annual Reviews, Palo Alto, 2013, Supplementary data (Supplementary Figure 1 - 8; Supplementary pp. 501–529. Movie 1 - 7; Supplementary Table 1) to this article including figure [9] M. Opekarová, J. Malinsky, W. Tanner, Plants and fungi in the era of hetero- geneous plasma membranes*, Plant Biol. 12 (2010) 94–98. captions can be found online at https://doi.org/10.1016/j.bbabio. [10] R.H. Pan, A.D. Jones, J.P. Hu, Cardiolipin-mediated mitochondrial dynamics and 2019.07.008. stress response in Arabidopsis, Plant Cell 26 (2014) 391–409. [11] K.M. Davies, M. Strauss, B. Daum, J.H. Kief, H.D. Osiewacz, A. Rycovska, V. Zickermann, W. Kuhlbrandt, Macromolecular organization of ATP synthase and Author contributions complex I in whole mitochondria, Proc. Natl. Acad. Sci. U. S. A. 108 (2011) 14121–14126. R.K. and J.K. designed the research. A.S., G.S., E.K. M.T. performed [12] J. Vonck, E. Schafer, Supramolecular organization of protein complexes in the the experiments; A.S. and M.T. generated PSI-YFP Synechocystis 6803 mitochondrial inner membrane, Biochim. Biophys. Acta Mol. Cell Res. 1793 (2009) 117–124. strain; A.S. and E.K. carried out physiological characterization of the [13] D.K. Schweppe, J.D. Chavez, C.F. Lee, A. Caudal, S.E. Kruse, R. Stuppard, new strain with YFP, G.S. was responsible for confocal microscopy D.J. Marcinek, G.S. Shadel, R. Tian, J.E. Bruce, Mitochondrial protein interactome measurements and G.K. together with G.S. then developed new elucidated by chemical cross-linking mass spectrometry, Proc. Natl. Acad. Sci. (2017). methods of image processing and statistical analysis of images. R.K., [14] V. Wilkens, W. Kohl, K. Busch, Restricted diffusion of OXPHOS complexes in dy- A.S. and J.K. wrote the article. All authors discussed the results and namic mitochondria delays their exchange between cristae and engenders a commented on the manuscript. transitory mosaic distribution, J. Cell Sci. 126 (2013) 103–116. [15] R. Acin-Perez, J.A. Enriquez, The function of the respiratory supercomplexes: the plasticity model, Biochim. Biophys. Acta, Bioenerg. 1837 (2014) 444–450. Transparency document [16] A.V. Ruban, M.P. Johnson, Visualizing the dynamic structure of the plant photo- synthetic membrane, Nat. Plants 1 (2015). [17] M. Pribil, M. Labs, D. Leister, Structure and dynamics of thylakoids in land plants, The Transparency document associated with this article can be J. Exp. Bot. 65 (2014) 1955–1972. found, in online version. [18] A. Mechela, S. Schwenkert, J. Soll, A brief history of thylakoid biogenesis, Open Biol. 9 (2019) 180237. [19] W.S. Chow, E.H. Kim, P. Horton, J.M. Anderson, Granal stacking of thylakoid Acknowledgments membranes in higher plant chloroplasts: the physicochemical forces at work and the functional consequences that ensue, Photochem. Photobiol. Sci. 4 (2005) The work was supported by project 16-10088S from the Czech 1081–1090.

13 A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053

[20] S. Puthiyaveetil, B. van Oort, H. Kirchhoff, Surface charge dynamics in photo- experimental protocols, Biochim. Biophys. Acta, Bioenerg. 1367 (1998) 88–106. synthetic membranes and the structural consequences, Nat. Plants 3 (2017) 9. [50] R.J. Porra, W.A. Thompson, P.E. Kriedemann, Determination of accurate extinc- [21] R. Kaňa, Govindjee, Role of ions in the regulation of light-harvesting, Front. Plant tion coefficients and simultaneous equations for assaying chlorophylls aandb Sci. 7 (2016). extracted with four different solvents: verification of the concentration of chlor- [22] P.-Å. Albertsson, A quantitative model of the domain structure of the photo- ophyll standards by atomic absorption spectroscopy, Biochim. Biophys. Acta synthetic membrane, Trends Plant Sci. 6 (2001) 349–354. Bioenergetics 975 (1989) 384–394. [23] N. Adir, Elucidation of the molecular structures of components of the phycobili- [51] M.D. Abramoff, P.J. Magelhaes, S.J. Ram, Image processing with ImageJ, some: reconstructing a giant, Photosynth. Res. 85 (2005) 15–32. Biophoton. Int. 11 (2004) 36–42. [24] H. Liu, H. Zhang, D.M. Niedzwiedzki, M. Prado, G. He, M.L. Gross, [52] T. Smith, J. Guild, The C.I.E. colorimetric standards and their use, Trans. Opt. Soc. R.E. Blankenship, Phycobilisomes supply excitations to both photosystems in a 33 (1931) 73. megacomplex in cyanobacteria, Science 342 (2013) 1104–1107. [53] H.S. Fairman, M.H. Brill, H. Hemmendinger, How the CIE 1931 color-matching [25] P.J. Keeling, The endosymbiotic origin, diversification and fate of , Philos. functions were derived from Wright-Guild data, Color. Res. Appl. 22 (1997) Trans. R. Soc., B 365 (2010) 729–748. 11–23. [26] V. Zimorski, C. Ku, W.F. Martin, S.B. Gould, Endosymbiotic theory for organelle [54] G. Konert, G. Steinbach, M. Canonico, R. Kaňa, Protein Arrangement (PA) factor: a origins, Curr. Opin. Microbiol. 22 (2014) 38–48. new photosynthetic parameter characterizing the organization of thylakoid [27] L.N. Liu, Distribution and dynamics of electron transport complexes in cyano- membrane proteins, Physiol. Plant. (2019) (accepted). bacterial thylakoid membranes, Biochim. Biophys. Acta, Bioenerg. 1857 (2016) [55] R.R. Sokal, F.J. Rohlf, Biometry: The Principles and Practice of Statistics in 256–265. Biological Research, Second ed., Freeman, New York, 1981. [28] A.M. Collins, M. Liberton, H.D.T. Jones, O.F. Garcia, H.B. Pakrasi, J.A. Timlin, [56] R. Kaňa, Mobility of photosynthetic proteins, Photosynth. Res. 116 (2013) Photosynthetic pigment localization and thylakoid membrane morphology are 465–479. altered in Synechocystis 6803 Phycobilisome mutants, Plant Physiol. 158 (2012) [57] R. Agarwal, A. Matros, M. Melzer, H.P. Mock, J.K. Sainis, Heterogeneity in thy- 1600–1609. lakoid membrane proteome of Synechocystis 6803, J. Proteome 73 (2010) [29] R. Nevo, D. Charuvi, E. Shimoni, R. Schwarz, A. Kaplan, I. Ohad, Z. Reich, 976–991. Thylakoid membrane perforations and connectivity enable intracellular traffic in [58] W.M. Elsasser, Outline of a theory of cellular heterogeneity, Proc. Natl. Acad. Sci. cyanobacteria, EMBO J. 26 (2007) 1467–1473. U. S. A. 81 (1984) 5126–5129. [30] M. Liberton, L.E. Page, W.B. O'Dell, H. O'Neill, E. Mamontov, V.S. Urban, [59] H. Rubin, Early origin and pervasiveness of cellular heterogeneity in some ma- H.B. Pakrasi, Organization and flexibility of cyanobacterial thylakoid membranes lignant transformations, Proc. Natl. Acad. Sci. U. S. A. 81 (1984) 5121–5125. examined by neutron scattering, J. Biol. Chem. 288 (2013) 3632–3640. [60] K.M. Davis, R.R. Isberg, Defining heterogeneity within bacterial populations via [31] W.F.J. Vermaas, J.A. Timlin, H.D.T. Jones, M.B. Sinclair, L.T. Nieman, single cell approaches, Bioessays 38 (2016) 782–790. S.W. Hamad, D.K. Melgaard, D.M. Haaland, In vivo hyperspectral confocal fluor- [61] S.J. Altschuler, L.F. Wu, Cellular heterogeneity: do differences make a difference? escence imaging to determine pigment localization and distribution in cyano- Cell 141 (2010) 559–563. bacterial cells, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 4050–4055. [62] C. Darwin, On the Origin of Species by Means of Natural Selection, or Preservation [32] D.M. Sherman, T.A. Troyan, L.A. Sherman, Localization of membrane-proteins in of Favoured Races in the Struggle for Life, London: John Murray, 1859, 1859. the cyanobacterium synechococcus sp pcc7942 - radial asymmetry in the photo- [63] M.B. Sinclair, D.M. Haaland, J.A. Timlin, H.D.T. Jones, Hyperspectral confocal synthetic complexes, Plant Physiol. 106 (1994) 251–262. microscope, Appl. Opt. 45 (2006) 6283–6291. [33] S. Casella, F. Huang, D. Mason, G.Y. Zhao, G.N. Johnson, C.W. Mullineaux, [64] A. Borodich, I. Rojdestvenski, M. Cottam, Lateral heterogeneity of photosystems in L.N. Liu, Dissecting the native architecture and dynamics of cyanobacterial pho- thylakoid membranes studied by Brownian dynamics simulations, Biophys. J. 85 tosynthetic machinery, Mol. Plant 10 (2017) 1434–1448. (2003) 774–789. [34] G. Steinbach, F. Schubert, R. Kaňa, Cryo-imaging of photosystems and phycobi- [65] N. Destainville, Cluster phases of membrane proteins, Phys. Rev. E 77 (2008) 5. lisomes in Anabaena sp. PCC 7120 cells, J. Photochem. Photobiol. B Biol. 152 [66] C.W. Mullineaux, Factors controlling the mobility of photosynthetic proteins, (2015) 395–399 Part B. Photochem. Photobiol. 84 (2008) 1310–1316. [35] C. MacGregor-Chatwin, M. Sener, S.F.H. Barnett, A. Hitchcock, M.C. Barnhart- [67] V.M. Sukhorukov, D. Dikov, K. Busch, V. Strecker, I. Wittig, J. Bereiter-Hahn, Dailey, K. Maghlaoui, J. Barber, J.A. Timlin, K. Schulten, C.N. Hunter, Lateral Determination of protein mobility in mitochondrial membranes of living cells, segregation of photosystem I in cyanobacterial thylakoids, Plant Cell 29 (2017) Biochim. Biophys. Acta, Biomembr. 1798 (2010) 2022–2032. 1119–1136. [68] T. Appelhans, K. Busch, Dynamic Imaging of Mitochondrial Membrane Proteins in [36] I.M. Folea, P. Zhang, E.M. Aro, E.J. Boekema, Domain organization of photosystem Specific Sub-Organelle Membrane Locations, (2017). II in membranes of the cyanobacterium Synechocystis PCC6803 investigated by [69] D. Ardail, J.P. Privat, M. Egretcharlier, C. Levrat, F. Lerme, P. Louisot, electron microscopy, FEBS Lett. 582 (2008) 1749–1754. Mitochondrial contact sites - lipid-composition and dynamics, J. Biol. Chem. 265 [37] S. Rexroth, C.W. Mullineaux, D. Ellinger, E. Sendtko, M. Rogner, F. Koenig, The (1990) 18797–18802. plasma membrane of the cyanobacterium Gloeobacter violaceus contains segre- [70] G. Hallermayer, W. Neupert, Lipid-composition of mitochondrial outer and inner gated bioenergetic domains, Plant Cell 23 (2011) 2379–2390. membranes of neurospora-crassa, Hoppe Seylers Z. Physiol. Chem. 355 (1974) [38] J. Mareš, P. Hrouzek, R. Kaňa, S. Ventura, O. Strunecký, J. Komárek, The primitive 279–288. thylakoid-less cyanobacterium Gloeobacter is a common rock-dwelling organism, [71] H. Kirchhoff, U. Mukherjee, H.J. Galla, Molecular architecture of the thylakoid PLoS One 8 (2013) e66323. membrane: lipid diffusion space for plastoquinone, Biochemistry 41 (2002) [39] G. Konert, G. Steinbach, M. Canonico, R. Kaňa, Protein arrangement factor: a new 4872–4882. photosynthetic parameter characterizing the organization of thylakoid membrane [72] H. Kirchhoff, Molecular crowding and order in photosynthetic membranes, Trends proteins, Physiol. Plant. 166 (2019) 264–277. Plant Sci. 13 (2008) 201–207. [40] A. Straskova, J. Knoppova, J. Komenda, Isolation of the cyanobacterial YFP-tagged [73] J.J. Sieber, K.I. Willig, C. Kutzner, C. Gerding-Reimers, B. Harke, G. Donnert, photosystem I using GFP-Trap (R), Photosynthetica 56 (2018) 300–305. B. Rammner, C. Eggeling, S.W. Hell, H. Grubmuller, T. Lang, Anatomy and dy- [41] J.G.K. Williams, Construction of specific mutations in photosystem-ii photo- namics of a supramolecular membrane protein cluster, Science 317 (2007) synthetic reaction center by genetic-engineering methods in Synechocystis-6803, 1072–1076. Methods Enzymol. 167 (1988) 766–778. [74] A. Kusumi, C. Nakada, K. Ritchie, K. Murase, K. Suzuki, H. Murakoshi, R.S. Kasai, [42] M. Tichy, M. Beckova, J. Kopecna, J. Noda, R. Sobotka, J. Komenda, Strain of J. Kondo, T. Fujiwara, Paradigm shift of the plasma membrane concept from the Synechocystis PCC 6803 with aberrant assembly of photosystem NN contains two-dimensional continuum fluid to the partitioned fluid: High-speed single-mo- tandem duplication of a large chromosomal region, Front. Plant Sci. 7 (2016). lecule tracking of membrane molecules, Annual Review of Biophysics and [43] T. Nagai, K. Ibata, E.S. Park, M. Kubota, K. Mikoshiba, A. Miyawaki, A variant of Biomolecular Structure, Annual Reviews, Palo Alto, 2005, pp. 351–U354. yellow fluorescent protein with fast and efficient maturation for cell-biological [75] M. Sarcina, N. Bouzovitis, C.W. Mullineaux, Mobilization of photosystem II in- applications, Nat. Biotechnol. 20 (2002) 87–90. duced by intense red light in the cyanobacterium Synechococcus sp PCC7942, [44] M. Dobakova, R. Sobotka, M. Tichy, J. Komenda, Psb28 protein is involved in the Plant Cell 18 (2006) 457–464. biogenesis of the photosystem II inner antenna CP47 (PsbB) in the cyanobacterium [76] G. Steinbach, R. Kaňa, Automated microscopy: macro language controlling a Synechocystis sp PCC 6803, Plant Physiol. 149 (2009) 1076–1086. confocal microscope and its external illumination – adaptation for photosynthetic [45] I. Wittig, M. Karas, H. Schagger, High resolution clear native electrophoresis for organisms, Microsc. Microanal. 22 (2016) 258–263. In-gel functional assays and fluorescence studies of membrane protein complexes, [77] M. Iwai, M. Yokono, A. Nakano, Visualizing structural dynamics of thylakoid Mol. Cell. Proteomics 6 (2007) 1215–1225. membranes, Sci. Rep. 4 (2014). [46] J. Komenda, J. Knoppova, J. Kopecna, R. Sobotka, P. Halada, J.F. Yu, J. Nickelsen, [78] D. Kirilovsky, R. Kaňa, O. Prášil, B. Demmig-Adams, G. Garab, W. Adams III, M. Boehm, P.J. Nixon, The Psb27 assembly factor binds to the CP43 complex of Mechanisms modulating energy arriving at reaction centers in cyanobacteria, in: photosystem II in the cyanobacterium Synechocystis sp PCC 6803, Plant Physiol. Govindjee (Ed.), Non-Photochemical Quenching and Energy Dissipation in Plants, 158 (2012) 476–486. Algae and Cyanobacteria, Springer Netherlands, 2014, pp. 471–501. [47] J. Knoppova, R. Sobotka, M. Tichy, J.F. Yu, P. Konik, P. Halada, P.J. Nixon, [79] S. Joshua, C.W. Mullineaux, Phycobilisome diffusion is required for light-state J. Komenda, Discovery of a chlorophyll binding protein complex involved in the transitions in cyanobacterial, Plant Physiol. 135 (2004) 2112–2119. early steps of photosystem II assembly in Synechocystis, Plant Cell 26 (2014) [80] C.W. Mullineaux, M.J. Tobin, G.R. Jones, Mobility of photosynthetic complexes in 1200–1212. thylakoid membranes, Nature 390 (1997) 421–424. [48] C. Klughammer, U. Schreiber, Saturation pulse method for assessment of energy [81] R. Kaňa, E. Kotabová, M. Lukeš, Š. Papáček, C. Matonoha, L.-N. Liu, O. Prášil, conversion in PS I, PAM Appl. Notes 1 (2008) 11–14. C.W. Mullineaux, Phycobilisome mobility and its role in the regulation of light [49] Z.S. Kolber, O. Prasil, P.G. Falkowski, Measurements of variable chlorophyll harvesting in red algae, Plant Physiol. 165 (2014) 1618–1631. fluorescence using fast repetition rate techniques: defining methodology and [82] M.D. McConnell, R. Koop, S. Vasil'ev, D. Bruce, Regulation of the distribution of

14 A. Strašková, et al. BBA - Bioenergetics 1860 (2019) 148053

chlorophyll and phycobilin-absorbed excitation energy in cyanobacteria. A grana stack, Trends Biochem. Sci. 18 (1993) 415–419. structure-based model for the light state transition, Plant Physiol. 130 (2002) [102] I.G. Tremmel, H. Kirchhoff, E. Weis, G.D. Farquhar, Dependence of plastoquinol 1201–1212. diffusion on the shape, size, and density of integral thylakoid proteins, Biochim. [83] K. El Bissati, E. Delphin, N. Murata, A.L. Etienne, D. Kirilovsky, Photosystem II Biophys. Acta, Bioenerg. 1607 (2003) 97–109. fluorescence quenching in the cyanobacterium Synechocystis PCC 6803: involve- [103] H. Kirchhoff, S. Horstmann, E. Weis, Control of the photosynthetic electron ment of two different mechanisms, Biochim. Biophys. Acta, Bioenerg. 1457 (2000) transport by PQ diffusion microdomains in thylakoids of higher plants, Biochim. 229–242. Biophys. Acta Bioenergetics 1459 (2000) 148–168. [84] W.M. Schluchter, G.H. Shen, J.D. Zhao, D.A. Bryant, Characterization of psaI and [104] J.M. Anderson, Consequences of spatial separation of photosystem-1 and photo- psaL mutants of Synechococcus sp strain PCC 7002: a new model for state tran- system-2 in thylakoid membranes of higher-plant chloroplasts, FEBS Lett. 124 sitions in cyanobacteria, Photochem. Photobiol. 64 (1996) 53–66. (1981) 1–10. [85] R. Kaňa, O. Prášil, O. Komárek, G.C. Papageorgiou, Govindjee, spectral char- [105] P. Joliot, J. Lavergne, D. Beal, Plastoquinone compartmentation in chloroplasts. 1. acteristic of fluorescence induction in a model cyanobacterium, Synechococcus sp Evidence for domains with different rates of photo-reduction, Biochim. Biophys. (PCC 7942), Biochim. Biophys. Acta 1787 (2009) 1170–1178. Acta 1101 (1992) 1–12. [86] R. Kaňa, E. Kotabová, O. Komárek, B. Šedivá, G.C. Papageorgiou, Govindjee, [106] J. Lavergne, J.P. Bouchaud, P. Joliot, Plastoquinone compartmentation in chlor- O. Prášil, The slow S to M fluorescence rise in cyanobacteria is due to a state 2to oplasts. 2. Theoretical aspects, Biochim. Biophys. Acta 1101 (1992) 13–22. state 1 transition, Biochim. Biophys. Acta Bioenergetics 1817 (2012) 1237–1247. [107] A. Magalon, F. Alberge, Distribution and dynamics of OXPHOS complexes in the [87] E. Tamary, V. Kiss, R. Nevo, Z. Adam, G. Bernat, S. Rexroth, M. Rogner, Z. Reich, bacterial cytoplasmic membrane, Biochim. Biophys. Acta Bioenergetics 1857 Structural and functional alterations of cyanobacterial phycobilisomes induced by (2016) 198–213. high-light stress, Biochim. Biophys. Acta, Bioenerg. 1817 (2012) 319–327. [108] J. Sacharz, S.J. Bryan, J. Yu, N.J. Burroughs, E.M. Spence, P.J. Nixon, [88] R. Ranjbar Choubeh, E. Wientjes, P.C. Struik, D. Kirilovsky, H. van Amerongen, C.W. Mullineaux, Sub-cellular location of FtsH proteases in the cyanobacterium State transitions in the cyanobacterium Synechococcus elongatus 7942 involve Synechocystis sp. PCC 6803 suggests localised PSII repair zones in the thylakoid reversible quenching of the photosystem II core, Biochim. Biophys. Acta membranes, Mol. Microbiol. 96 (2015) 448–462. Bioenergetics 1859 (2018) 1059–1066. [109] L.-N. Liu, S.J. Bryan, F. Huang, J. Yu, P.J. Nixon, P.R. Rich, C.W. Mullineaux, [89] K.B. Busch, G. Deckers-Hebestreit, G.T. Hanke, A.Y. Mulkidjanian, Dynamics of Control of electron transport routes through redox-regulated redistribution of re- bioenergetic microcompartments, Biol. Chem. 394 (2013) 163–188. spiratory complexes, Proc. Natl. Acad. Sci. U. S. A. 109 (2012) 11431–11436. [90] T. Mirkovic, E.E. Ostroumov, J.M. Anna, R. van Grondelle, Govindjee, [110] S. Heinz, A. Rast, L. Shao, A. Gutu, I.L. Gugel, E. Heyno, M. Labs, B. Rengstl, G.D. Scholes, Light absorption and energy transfer in the antenna complexes of S. Viola, M.M. Nowaczyk, D. Leister, J. Nickelsen, Thylakoid membrane archi- photosynthetic organisms, Chem. Rev. (2016). tecture in Synechocystis depends on CurT, a homolog of the Granal CURVATURE [91] E. Crisafi, A. Pandit, Disentangling protein and lipid interactions that controla THYLAKOID1 proteins, Plant Cell 28 (2016) 2238–2260. molecular switch in photosynthetic light harvesting, Biochim. Biophys. Acta, [111] M. Schottkowski, J. Ratke, U. Oster, M. Nowaczyk, J. Nickelsen, Pitt, a novel Biomembr. 1859 (2017) 40–47. tetratricopeptide repeat protein involved in light-dependent chlorophyll bio- [92] S. Haferkamp, W. Haase, A.A. Pascal, H. van Amerongen, H. Kirchhoff, Efficient synthesis and thylakoid membrane biogenesis in Synechocystis sp PCC 6803, Mol. light harvesting by photosystem II requires an optimized protein packing density Plant 2 (2009) 1289–1297. in grana thylakoids, J. Biol. Chem. 285 (2010) 17020–17028. [112] A. Rast, S. Heinz, J. Nickelsen, Biogenesis of thylakoid membranes, Biochim. [93] S. Caffarri, R. Kouril, S. Kereiche, E.J. Boekema, R. Croce, Functional architecture Biophys. Acta, Bioenerg. 1847 (2015) 821–830. of higher plant photosystem II supercomplexes, EMBO J. 28 (2009) 3052–3063. [113] M. Bečková, Z. Gardian, J. Yu, P. Konik, P.J. Nixon, J. Komenda, Association of [94] P. Galka, S. Santabarbara, T.T.H. Khuong, H. Degand, P. Morsomme, Psb28 and Psb27 proteins with PSII-PSI supercomplexes upon exposure of R.C. Jennings, E.J. Boekema, S. Caffarria, Functional analyses of the plant pho- Synechocystis sp. PCC 6803 to high light, Mol. Plant 10 (2017) 62–72. tosystem I-light-harvesting complex II supercomplex reveal that light-harvesting [114] B. Andersson, J.M. Anderson, Lateral heterogeneity in the distribution of chlor- complex II loosely bound to is a very efficient antenna for photo- ophyll-protein complexes of the thylakoid membranes of spinach-chloroplasts, system I in state II, Plant Cell 24 (2012) 2963–2978. Biochim. Biophys. Acta 593 (1980) 427–440. [95] M. Rantala, M. Tikkanen, E.M. Aro, Proteomic characterization of hierarchical [115] D. Kaftan, V. Brumfeld, R. Nevo, A. Scherz, Z. Reich, From chloroplasts to pho- megacomplex formation in Arabidopsis thylakoid membrane, Plant J. 92 (2017) tosystems: in situ scanning force microscopy on intact thylakoid membranes, 951–962. EMBO J. 21 (2002) 6146–6153. [96] M. Yokono, A. Takabayashi, S. Akimoto, A. Tanaka, A megacomplex composed of [116] L. Mustardy, G. Garab, Granum revisited. A three-dimensional model - where both photosystem reaction centres in higher plants, Nat. Commun. 6 (2015) 6675. things fall into place, Trends Plant Sci. 8 (2003) 117–122. [97] H. Kirchhoff, Structure-function relationships in photosynthetic membranes: [117] B.E. Pfeil, B. Schoefs, C. Spetea, Function and evolution of channels and trans- challenges and emerging fields, Plant Sci. 266 (2018) 76–82. porters in photosynthetic membranes, Cell. Mol. Life Sci. 71 (2014) 979–998. [98] H. Kirchhoff, Diffusion of molecules and macromolecules in thylakoid membranes, [118] S. Govindarajan, O. Amster-Choder, Where are things inside a bacterial cell? Curr. Biochim. Biophys. Acta, Bioenerg. 1837 (2014) 495–502. Opin. Microbiol. 33 (2016) 83–90. [99] J.P. Dekker, E.J. Boekema, Supramolecular organization of thylakoid membrane [119] P.M. Llopis, A.F. Jackson, O. Sliusarenko, I. Surovtsev, J. Heinritz, T. Emonet, proteins in green plants, Biochim. Biophys. Acta, Bioenerg. 1706 (2005) 12–39. C. Jacobs-Wagner, Spatial organization of the flow of genetic information in [100] R. Kouril, J.P. Dekker, E.J. Boekema, Supramolecular organization of photosystem bacteria, Nature 466 (2010) 77–81. II in green plants, Biochim. Biophys. Acta, Bioenerg. 1817 (2012) 2–12. [120] J.R. Maddock, L. Shapiro, Polar location of the chemoreceptor complex in the [101] H.W. Trissl, C. Wilhelm, Why do thylakoid membranes from higher-plants form Escherichia-coli cell, Science 259 (1993) 1717–1723.

15